Inorg. Chem. 1982, 21, 3336-3341
3336
Contribution from the Department of Chemistry, University of New Mexico, Albuquerque, New Mexico 87131
Stereochemistry of Arsenic(II1) and Antimony(II1) 1,2-Dihydroxycyclohexane-1,2-dicarboxylates DRAGOSLAV MARCOVICH, EILEEN N. DUESLER, ROBERT E. TAPSCOTT,* and THEODORE F. THEM Received March 22, 1982
Solid salts of complexes of arsenic(II1) and antimony(II1)with (&)- and meso- 1,2-dihydroxycyclohexane-1,2-dicarboxylate(4-) (dhcdc) having the general formulas Na2[M2(CsHs06)2].xH20 have been prepared and characterized. I3CNMR spectroscopy shows that, with the probable exception of the arsenic(II1) ms-dhcdc complex, the complexes are binuclear in solution though the arsenic(II1) (A)-dhcdcspecies is extensively dissociated. The presence of a well-defined @-AA(RS,RS) and @-AA(RS,RS) enantiomeric pair of C2-symmetryisomers for the antimony(II1) ms-dhcdc complex in aqueous solution is indicated by 13CNMR spectroscopy. The crystal structure of Na2[A~2((&)-CsH806)2].6H20 (P2,/c, a = 9.128 (2) A, 6 = 16.867 (4) A, c = 16.039 (4) A, p = 91.04 (2)O, 2 = 4) shows the presence of a racemic mixture of binuclear AA(RR,RR) and AA(SS,Ss) isomeric complexes in the solid and shows indication of serious strain induced in the ligand by complex formation. This is the first report of complexes containing cyclic tartrate derivatives as ligands.
Introduction A recent paper' reported 13C N M R studies of binuclear arsenic(II1) and antimony(II1) dihydroxydicarboxylate(4-) complexes (1) in aqueous solution. Besides confirming the 2-
cyclohexanedicarboxylic acid show only ee structure^.^,^ However, it is precisely the higher energy aa conformer that is required for bridging in binuclear structures containing (&)-dhcdc. Moreover, molecular models indicate that, in such a complex, the axial carboxyl groups will be turned such that carboxyl oxygen atoms are in close contact with syn axial hydrogen atoms-a distinctly unfavorable conformation (2).*
1
binuclear structures, these studies showed (1) the predicted* greater stability for isomeric complexes containing two bridges of the same enantiomeric form (RR,RR and SSJS) compared to those containing bridges of opposite chirality (RR,SS),(2) the unexpected3 existence of meso-ligand-containing binuclear complexes of arsenic(III), (3) unusually large long-range 13C-lH coupling constants, and (4) large 13Cchemical shift changes accompanying formation of the mixed As/Sb complexes-changes that may be due to bond angle deformations. In this paper, we report synthetic, I3C N M R spectroscopic, and X-ray diffraction studies of complexes of arsenic(II1) and antimony(II1) with 1,2-dihydroxycyclohexane-1,2-dicarboxylate (dhcdc). This ligand, which is the first cyclic tartrate derivative to be employed in complexation, is restricted in its conformational freedom compared to the acyclic tartrate derivatives studied earlier.' For (*)-dhcdc (the trans or RR,SS isomer), two conformers-axial, axial or aa and equatorial, equatorial or e e a r e possible, while for the meso (cis or R S ) isomer, the two possible conformers are enantiomeric and are ae (Figure 1). Here we have designated the conformer according to the position of the carboxyl group. The aa conformer of (*)-dhcdc is expected to be less stable than the ee conformer from two considerations. First, axial/equatorial energy differences for hydroxyl4 and carboxyl5 groups allow a rough calculation of 2 kcal mol-' free energy difference between the two conformers if the substituents are unionized and experience no mutual nonbonded interactions. Second, X-ray structure results for the related (k)-1,2(1) Marcovich, D.; Tapscott, R. E. J . Am. Chem. SOC.1980, 102, 5712. (2) Tapscott, R. E. Inorg. Chim. Acra 1974, 10, 183. (3) Tapscott, R. E.; Belford, R. L.; Paul, I. C. Coord. Chem. Reo. 1969, 4, 323. (4) Eliel, E. L. J . Chem. Educ. 1960, 37, 126. (5) Stolow, R. D. J . Am. Chem. SOC.1959, 81, 5806.
2
On the other hand, the meso isomer would bridge by precisely those conformers found for the free ligand. Thus we might expect a destabilization of the complex formed by the racemic ligand and, possibly, a stabilization of the complex formed by the meso ligand. Indeed, work with vanadyl(1V) has already shown that (*)-dhcdc is unusual in its behavior in apparently complex formation in that, though a binuclear complex can be obtained, it cannot be deprotonated completely to give a tetranegative anionic specie^.^ This behavior is unlike that found for any other vanadyl(1V) dihydroxydicarboxylate investigated to date.1° Experimental Section 1,2-Dihydroxycyclohexane-1,2-dicarboxylicAcids. The crude mixture of isomeric ligands was prepared by hydrolysis of the cyanohydrins of 1,2-cyclohexanedionefollowing the procedures of Tatsumi et al.;" however, to avoid the ester preparation and distillation employed previously, we used repeated precipitations and dissolutions of the barium salts for product purification. In addition, barium salts rather than calcium salts were employed for the isomer separations. The former proved easier to work with. These modifications are described below.
The orange-brown hydrolysate obtained by hydrolysis of the cyanohydrins prepared" from 25 g of 1,2-cyclohexanedionewas suction filtered to remove tars and large dark crystals, was made basic (pH 10) with 15 M NH,OH, at which time the solution (volume, 1
-
-
(6) Benedetti, E.; Corradini, P.; Pedone, C.; Post, B. J . Am. Chem. Soc. 1969, 91, 4072. (7) Benedetti, E.; Corradini, P.; Pedone, C. J . Am. Chem. SOC.1969, 91, 4075. (8) Allinger, N. L.; Chang, S. H. M. Tetrahedron 1977, 33, 1561. (9) Them, T. F. Ph.D. Dissertation, University of New Mexico, 1977. (10) Hahs, S. K.; Ortega, R. B.; Tapscott, R. E.; Campana, C. F.; Morosin, B. Inorg. Chem. 1982, 21, 664. ( 1 1 ) Tatsumi, S . ; Imaida, M.; Izumi, Y . Bull. Chem. SOC.Jpn. 1966, 39, 2543.
0020-1669/82/ 1321-3336$01.25/0 0 1982 American Chemical Society
(M = As(III), Sb(II1))
O(H)
Figure 1. Conformers of 1,2-dihydroxycyclohexane-1,2-dicarboxylic acid and its anions. The trans S,S diastereomer is shown at the top, and the cis R,S (meso) diastereomer, at the bottom. L) became turbid, and was decolorized with charcoal. To the clear yellow filtrate was added a solution of 55 g of BaCI2.2H20 dissolved in 300 mL of HzO. The mixture was stirred for 2 h, and the creamy white precipitate of the barium salt of dhcdc was collected by suction filtration and washed with hot water. The wet precipitate was slurried with 100 mL of water, and 12 M HCI was added dropwise while it was heated on a steam bath to effect dissolution. The yellow-brown solution was filtered and made basic (pH 10) with N H 4 0 H , and additional BaC12.2H20 solution (1 g in 5 mL of water) was added. After being stirred for 5 min, the precipitate was collected, washed with hot water, and subjected to this dissolution/precipitation process twice more. The final filtered precipitate was washed with hot water and diethyl ether and was dried with suction for 12 h to give 25.5 g of a white powder. Anal. Calcd for BaC&&,.1'/zH20: C, 26.22; H, 3.58; Ba, 37.47. Found: C, 26.31; H , 3.56; Ba, 37.00. For the separation of the meso and racemic forms, a slurry of 10.0 g of the barium salts in 500 mL of water was treated with 12 M HCI, added dropwise with stirring, to effect dissolution (final volume, 550 mL). Upon adjustment of the p H to -4.3 with 15 M N H 4 0 H , immediate precipitation of the crude barium salt of the meso isomer occurred. The precipitate was collected, washed with hot water and diethyl ether, and suction dried to give 3.6 g of white powder. The filtrate was retained for isolation of the racemic compound. Since the IR spectrum indicated a slight contamination by racemic material, the barium salt of ms-dhcdc was slurried with 250 mL of water and allowed to react with 50 mL of 12 M HC1 while it was heated and stirred. The solution was filtered and titrated to pH 3.9 with 60 mL of 15 M NHIOH to give an immediate precipitate of the pure meso salt. The product was collected, washed with diethyl ether, and dried with suction to give 1.7 g of white powder. Anal. Cakd for BaC8H1006~11/2H20: c , 26.22; H, 3.58; Ba, 37.47. Found: C, 26.01; H, 3.60; Ba, 37.21. I R (KBr) peaks characteristic of Ba(ms-dhcdc): 1002, 1128, 1257 cm-l. The racemic material was isolated from the pH 4.3 filtrate from which the crude barium salt of the meso isomer was obtained. The filtrate was made basic to pH 10 with 15 M NH40H, and the barium salt of the racemic isomer precipitated immediately and was collected after the volume had been reduced to 100 mL on a rotary evaporator. The product was washed with hot water, dissolved in 25 mL of water with dropwise addition of 12 M HCI, and reprecipitated from the filtered solution by addition of 15 M N H 4 0 H to give pH 10. The barium salt of the racemic ligand (with slight contamination by the meso compound) was collected, washed with hot water and diethyl ether, and dried with suction to give 3.5 g of white powder. Anal. Calcd for BaC8Hlo06.11/2H20:C, 26.22; H, 3.58; Ba, 37.47. Found: C, 24.74; H, 3.69; Ba, 37.10. I R (KBr) peak characteristic of the racemic compound: 1109 cm-I. The free acids were prepared by dissolution of the barium salts in 1 M HC1, evaporation of the solution to dryness on a rotary evaporator, and extraction of the residue with acetone. Reduction in volume of the combined acetone extracts and addition of a mixture of ligroin and CC14 precipitated the acids, which were recrystallized from hot glacial acetic acid. Satisfactory elemental analyses were obtained, and the melting points were in excellent agreement with literature values.5 The "C N M R band positions are reported in Table 111. IR (KBr): peaks characteristic of (f)-C8HI2O6702, 743, 910, 978, 1062 cm-'; peaks characteristic of ms-C8Hlz06751, 800, 962, 1091, 1151 cm-'. Disodium salts of the 1,2-dihydroxycyclohexane1,2-dicarboxyIic acid isomers were prepared by evaporation of aqueous solutions of the appropriate acid neutralized with NaOH. Anal. Calcd for
Inorganic Chemistry, Vol. 21, No. 9, 1982 3331 (f)-Na2C8H@6*H20: c , 36.10; H, 4.54. Found: c , 36.63; H, 4.28. IR (KBr): peaks characteristic of the racemic salt 899, 1025, 1073, 1174 cm-I. Anal. Calcd for mS'-Na&Hlo06: c , 38.44; H, 4.06. Found: C, 38.44; H, 4.20. IR (KBr): peaks characteristic of the meso salt 659, 931, 1095, 1138 cm-I. Sodium (1,2-Dihydroxycyclohexane-1,2-dicarboxylato)arte(III) and -antimonate(III). All of the complex salts prepared lost water easily to give variable degrees of hydration depending on how the compounds were dried. The analyses given are for the product with only short drying times under ambient conditions and presumably reflect the degree of hydration as initially obtained. Anhydrous compounds could not be obtained without decomposition. The complex salts exhibited significant temperature dependencies of solubility in water, and recrystallizations from this solvent were easily effected. Recrystallizations from ethanol/water mixtures often gave products containing variable amounts of ethanol so that this technique was generally not used. Sodium salts of the arsenic(II1) complex were prepared by mixing cooled aqueous solutions of sodium arsenite, NaAs02, and the appropriate 1,2-dihydroxycyclohexane1,2-dicarboxylic acid in a 1:1 mole ratio (caution!exothermic reaction), filtering the solution, allowing the solvent to evaporate, and collecting the crystals. Anal. Calcd for Na2[As2((f)-C8H806)2].6Hz0: c , 27.29; H, 4.06. Found: c, 27.40; H, 4.06. Cdcd for Na2[As2(ms-C8H80,),].51/2H20: c , 27.65; H, 3.91. Found: C, 27.78; H, 3.86. Drying in vacuo for 1-3 days at 65-80 "C and for 4 days at 150 O C gave complex salts prepared from the racemic ligand that were found to agree very well with analyses as a 4-hydrate and as a 2-hydrate, respectively. The salt prepared from the meso ligand reproducibly gave a 5-hydrate upon further drying under ambient conditions. Sodium salts of the antimony(II1) complexes were prepared from antimony(II1) oxide, Sb2O3, and the appropriate 1,2-dihydroxycyclohexane- 1,2-dicarboxyIic acid with the procedure described elsewhere for other dihydroxydicarboxylate complexes,' except that ethanol was not used for the isolations of the solids. Two to four days of heating and an excess (-25%) of Sb2O3 were required for the best yields. The excess Sb2O3 was removed by filtration. Anal. Calcd for Na2[Sb2((*)-C8H8O6)z].5H20:C, 24.64; H , 3.36. Found: C, 24.25; H, 3.17. Calcd for Na2[Sb2(ms-C,H,06)2].7H20: C, 23.56; H, 3.71. Found: C, 23.77; H, 3.84. Physical Measurements. Natural-abundance I3C N M R spectra were recorded at ca. 40 "C on a Varian FT-80 N M R spectrometer with an observation frequency of 20.0 MHz. 13C N M R spectra of the free ligands were obtained with use of a 5000-Hz sweep width, 0.8-s acquisition time, 10-ps pulse width (50" pulse), and a 0.5-s pulse delay. Spectra of the complexes were obtained with use of a 500-Hz sweep width, 1.1-s acquisition time, 7-ps pulse width (40" pulse), and no pulse delay. For the N M R samples, material was dissolved in 2.0 mL of D 2 0 and placed in a 10" sample tube. Tetramethylsilane was used as an external standard in a coaxial 5-mm tube. Crystal Structure Determination. A single crystal of Na2[As2((f)-C8H806)2]'6H20 was grown by cooling a solution of the compound in ethanol/water and was mounted in a glass capillary. Precession photographs established the space group as monoclinic, P 2 , / c , with systematic absences h01, I = 2n 1, and OkO, k = 2n + 1. Diffraction data were collected on a Syntex P3/F diffractometer system at ambient temperature (-23 "C) and were treated for Lorentz and polarization effects. An empirical absorption correction based on $ scans was applied. The crystal and data collection parameters are tabulated in Table I. Neutral-atom scattering factors and anomalous dispersion corrections were used for all atoms during the refinements.12 An origin-removed, sharpened Patterson map was interpreted for two trial arsenic atom positions. From Fourier maps, phased initially on these arsenic atom positions, and difference Fourier maps, the positions of all nonhydrogen atoms were obtained. Refinement of these atom positions using isotropic thermal parameters converged a t R = 10%. The ring hydrogen atoms were calculated in idealized positions, and their isotropic thermal parameters were set at 1.2 times the isotropic thermal parameter of their parent carbon atoms. Fourier
+
were done with the SHELXTL 79 package: Sheldrick, G . M. 'Nicolet SHELXTL Operations Manual"; Nicolet XRD Corp.: Cupertino, CA, 1979. SHELXTL uses scattering factor data compiled in: "International Tables for X-ray Crystallography"; Kynoch Press: Birmingham, England, 1968; Vol. IV, pp 55, 99, 149.
(12) All calculations and plots
3338 Inorganic Chemistry, Vol. 21, No. 9, 1982 Table I. Summary of Crystal and Data Collection Parameters for Na, [As,(C,H ,O, l2 1 .6H20
v,.A3
z
p(calcd),g space group cryst dimens, mm radiation monochromator
617.3 1424 9.128 (2) 16.867 (4) 16.039 (4) 91.04 (2) 2469 (2) 4 1.89 p2 1I C 0.2 X 0.35 X 0.55 Mo Kcr (A = 0.7 10 69 A)
graphite
abs coeff, ~ ( M oKcr), mP
29.9
scan speed, deg min-' scan range
variable, 4.0-30.0
1.0" below 20 ( K c Y ~to) 1.3" above 2e (Kol,) 28 limits, deg 1.0-60.0 bkgd counting time 0.5 x (scan time) 2 stds every 94 reflctns stds monitored 7891 reflctns collected unique reflctns 7234 5085 with F z > 2 0 ( p ) ~ m , unique data used n , no. of variables refined 344 E R F = [ Z W ( I F ~ I IFcl)Z/(m-n)]'R 1.14 W-1 o2(F0) + (gF,)*,g = 0.032 R(F) = Z I IFo I - lFcll/Z lFol 0.045 R w ( F ) = [ L : ~ ( I F o l -1FCl)*/ 0.047 c w IFoI,] ", R ( F ) including all data 0.070 0.00012 x, extinction parameter in F = Fo(l- 10-4xFoZ/sin e) a
Six bad reflections (11,2,10; 11,3,-6; 11,3,0; 11,3,3; 11,3,9;
11,4,-5) were deleted.
Marcovich et al. Table 11. Fractional Coordinates for Nonhydrogen Atoms in Na, [As, (C& 0 6 ), 1.6H,O atom X Y Z 0.44050 (4) 0.08322 (4) 0.7414 (2) 0.3205 (2) 0.3018 (3) 0.0661 (3) 0.2791 (3) 0.0602 (3) 0.0726 (3) 0.0502 (4) 0.5232 (3) 0.5689 (3) 0.4343 (3) 0.2750 (3) 0.1528 (3) 0.3583 (4) 0.4738 (3) 0.1327 (3) 0.3055 (4) 0.2324 (4) 0.0426 (3) 0.3470 (3) 0.1651 (4) 0.1439 (4) 0.0307 (4) 0.0529 (4) 0.5323 (4) 0.4955 (4) 0.3764 (4) 0.2944 (4) 0.0986 (4) -0.0594 (4) -0.1641 (4) -0.1289 (4) 0.6346 (4) 0.6917 (5) 0.5742 (5) 0.4384 (4)
0.19727 (2) 0.29743 (2) 0.4419 (1) 0.3416 (1) 0.2476 (2) 0.2541 (2) 0.1481 (1) 0.2688 (1) 0.1778 (2) 0.0647 (2) 0.1638 (1) 0.2099 (2) 0.2932 (2) 0.2812 (1) 0.3981 (2) 0.4592 (2) 0.4096 (2) 0.4368 (2) 0.5506 (2) 0.6943 (2) 0.5838 (2) 0.4568 (2) 0.2267 (2) 0.1610 (2) 0.1874 (2) 0.1374 (2) 0.2202 (2) 0.3029 (2) 0.3422 (2) 0.4054 (2) 0.0842 (2) 0.0838 (3) 0.1002 (2) 0.1794 (2) 0.3535 (2) 0.3850 (3) 0.4318 (3) 0.3799 (2)
0.35729 (2) 0.16548 (2) 0.4521 (1) 0.5510 (1) 0.4368 (2) 0.4657 (2) 0.3187 (2) 0.2728 (1) 0.1460 (2) 0.2152 (2) 0.2467 (2) 0.1206 (2) 0.3082 (2) 0.1466 (2) 0.2161 (2) 0.2562 (2) 0.4450 (2) 0.5145 (2) 0.3980 (2) 0.5168 (2) 0.6329 (2) 0.65 16 (2) 0.4252 (2) 0.3599 (2) 0.2918 (2) 0.2131 (2) 0.1932 (2) 0.2273 (2) 0.1712 (2) 0.2204 (2) 0.4040 (3) 0.4327 (3) 0.3600 (3) 0.3189 (2) 0.2358 (2) 0.1530 (3) 0.1069 (3) 0.0917 (2)
difference maps determined after anisotropic refinement of all nonhydrogen atoms yielded reasonable positions for the water molecule hydrogen atoms. These were included in the final refinement with positions fixed at those found in the difference maps and with a common invariant isotropic thermal parameter of U = 0.05. A final difference map showed the highest four peaks, which were all less than 0.9 %, from the arsenic atoms, to correspond to electron densities of 0.90-0.52 e A-3. The remaining peak heights were less than 0.43 e A-3. Final agreement factors are tabulated in Table I, and the final fractional coordinates are listed in Table 11. Tables of anisotropic thermal parameters, hydrogen atom parameters, hydrogen bonding distances and angles, sodium ion coordination geometries, best planes, and observed and calculated structure factors are available as supplementary material. Results and Discussion Ligand Stereochemistry. Chemical shifts obtained from aqueous solution I3C N M R spectra of the diastereomeric forms of dhcdc, both as the free acids and as the dinegative anions, are presented in Table 111. Here we use these data to support our earlier prediction that the lowest energy conformer of the racemic isomer of dhcdc is ee (remember that this designation gives the positions of the carboxyl groups only). I3C N M R spectral data that have been reported for 1,2cyclohexanedicarboxylic acidI3 show carboxyl, P-CH2, and y-CH, chemical shifts which are at, respectively, 1.2, 2.6,and 1.4 ppm higher field for the conformationally averaged meso diastereomer than for the racemic. Shifts to higher field for these carbon resonances are precisely what is expected if the racemic isomer is ee and if the axial carboxyl group of the (meso) ae conformer gives rise to a y steric shift owing to 1,3-diaxial interactions with methylene groups.14 The I3C
N M R spectra of the diastereomers of dhcdc, on the other hand, show (Table 111) methylene resonance position differences between the meso and racemic isomers that are reversed from those found for 1,2-cyclohexanedicarboxylicacid. The meso isomer resonances are downfield from those of the racemic species. Though this might seem to indicate a conformational difference between 1,2-cyclohexanedicarboxylic acid diasteromers and their 1,2-dihydroxy derivatives, in fact it is what would be predicted. An ae conformer for ms-dhcdc will have one axial carboxyl group and one axial hydroxyl group. On the other hand, an ee conformer of (A)-dhcdc will have two axial hydroxyl groups. Thus, the methylene resonances of the racemic isomer are expected to occur at higher field than those of the conformationally averaged meso isomer if axial hydroxyl groups give rise to larger y shifts than do axial carboxyl groups, and this is believed to be the case. Not only are axial hydroxyl groups in cyclohexanols known to give large upfield shifts to y methylene carbonsI6 but hydroxyl groups are also known to give upfield y shifts in general acyclic compounds that are about 1.7 ppm larger than those found for methyl groupsI6 (which are comparable to carboxyl group^"^'^). Moreover, the carboxyl resonance of ms-dhcdc is upfield from that of the racemic diastereomer. This is the same ordering as found for 1,2-cyclohexanedicarboxylicacid1 and is the ordering expected for an upfield y effect on the more axial carboxyl groups of the meso diastereomers. Finally, we
(13) Williamson, K. L.; Hasan, M. U.; Clutter, D. R. J . Magn. Reson. 1978, 30, 367. (14) Though the origin of 'steric" y shifts is highly controversial, there is no doubt that a generalized heteroatom y effect has been found experimentally for a variety of substituent^.'^
( 1 5 ) Seidman, K.; Maciel, G. E. J . Am. Chem. SOC.1977, 99, 659. (16) Roberts, J. D.; Weigert, F. J.; Kroschwitz, .I.I.; Reich, H. J. J . Am. Chem. SOC.1970, 92, 1338. (17) Grant, D. M.; Paul, E. G. J . Am. Chem. SOC.1964, 86, 2984. (18) MacDonald, J. C. J . Magn. Reson. 1979, 34, 207.
Inorganic Chemistry, Vol. 21, No. 9, 1982 3339
Na2[MZ(CsHs06)z1~xHz0 (M = As(III), Sb(II1))
Table 111. "C NMR Chemical Shifts far Isomers of 1,2-Dihydroxycyclohexane-1,2-dicarboxyliAcid and Its Compounds 6
compd
wncn, M
-COO(H) 178.85 [1.0] 177.89 [l.O] 184.70 [LO] 181.82 [1.0] 181.43 [l.O]
>CO(H)
[amplitude]' EH,
rCH,
77.57 [1.3] 31.82 [2.9] 20.39 [3.2] 78.42 [1.3] 33.28 [5.6] 21.02 [6.6] 1.0 77.69 [ 2 . 2 ] 32.15 [2.2] 20.90 [2.1] 1.0 79.81 [1.4] 33.70 [5.0] 22.32 [6.3] 0.2 83.97 [1.4] 36.22 [2.6] 22.30 [2.61 77.28 [0.SIb 31.98 [1.3]* 20.61 [2.81b 0.2 184.34 83.03 [2.0] 22.85 [7.5] 37.42 [6.51 0.2 80.1d 180.gd 34.34 21.Y 0.2 83.92 [1.4lC 184.21 [1.0lC 36.41 [0.8lC 22.68 [1.1]' 82.42 [1.6lC 184.07 [1.3Ie 20.20 [0.9lC 33.62 (0.81' a Peak height relative to that of the peak at lowest field. These resonances are not assigned to the binuclear dhcdc complex. They are probably due to partially or totally dissociated ligand. e Srjghtly broadened. Severely broadened (1-3 ppm at half-height);resonance Center piven. 1.0 1.0
P-AAIRSRS
/3-Ah(RS.RS)
Figure 2. Isomers expected for the complexes [M,((*)-dhcdc),]* and [M,(ms-dhcdc),l2-, M = Sb,As.
note that ionized carboxyl groups give rise to decreased y shifts18and this should lead to decreases in the chemical shift differences between corresponding methylene carbon atoms of the meso and racemic diastereomers of dhcdc upon ionization. This is precisely what is seen (Table 111). All of these observations provide strong evidence for our conclusion of an ee preferred conformer for free (f)-dhcdc. ''C NMR Spectra of Complexes. The isomers expected for binuclear complexes containing metal ions with pseudo-trigonal-bipyramidal coordination geometries (as found in the ar~enic(II1)'~ and antimony(II1)' tartrates) are the enantiomeric pairs AA(RR,RR), AA(SS,SS) for the racemic system and @-AA(RSW), P-AA(RSW) for the meso (Figure 2)?." The former pair should have D, symmetry with four "C N M R resonances expected-one each from carboxyl, carbinol, and two types of methylene carbon atoms ((3 and y to the carboxyl groups). The latter pair of isomers is expected to have C, symmetry with two resonances for each type of carbon atom (eight resonances total). The spectra predicted are precisely those observed for the antimony(II1) complexes (Table 111). In these complexes, the large downfield shifts of the carbinol resonances upon complexation show conclusively ionization and coordination of the hydroxyl groups,'~z'~zz as expected. The downfield shifts of the methylene carbon resonances in the coordinated racemic (19) Matsumoto. K.; Kawaguchi. H.; Kuroya. Chem. Soe. Jpn. 1973.46, 2424.
(20)
H.; Kawaguchi, S. Bull.
These designations differ from thmc of ref 2, which gives the isomers as AA(d4. W I O and B A A ( m - m ) . @-AA(ms-ms). The change was
made in order to differentiateby abaolute mnfiguration rather than by sign of optical rotation. (21) . . Bera. LE.: Brandbee. - . S.: . Lindblam.. L.:. Werner. P.-E. Acta Chem. ScGd.. Ser. A 1977, A31, 325. (22) Tselinskii, Yu.K.; Pcstryakw. B. V.; Olivson, A. I.; Pekhterm. T. M. Zh. Neorg. Khim. 1979, 24, 1243.
ligand compared to those of the free ligand are undoubtedly due to a decreased y perturbation since the more perturbing hydroxyl groups must be equatorial in the AA(RR,RR), AA(SS,SS)complexes. Interestingly, the antimony(II1) &AA(RS,RS), p-AA(RS,RS)complex shows one &CH, resonance (36.41 ppm) and one y-CH, resonance (36.41 ppm) which are very close to those determined for the (*)-dhcdc complex. These resonances must correspond to methylene groups y to the axial carboxyl group. The other methylene resonances (33.62, 20.20 ppm) of the RS,RS isomers are upfield and correspond to methylene groups y to the axial hydroxyl group. Note that, in the coordinated mr-dhcdc ligand, conformational averaging does not occur to give NMR equivalency to carbon atoms of the same type (i.e., same connectivity), as is the case with the free ligand. The absence of any peaks attributable to a coexisting isomer in solution indicates the instability of binuclear complexes with bridging ligands of opposite chirality, RR,SS, relative to that of the RR,RR and SS,SS species.' The 'C N M R spectra of the arsenic(II1) complexes (Table 111) are not as simple as those of the antimony(II1) complexes. In the spectrum of the arsenic(II1) complex of (+)-dhcdc, in addition to peaks a t 181.43, 83.97, 36.22, and 22.30 ppm, which we attribute to binuclear AA(RR,RR), AA(SS,SS) complexes, extra resonances are found. Similar additional resonances are observed in spectra of arsenic(II1) tartrates and substituted tartrates.' If one extra resonance peak is overlapped with the peak at 181.43 ppm, all of the extra resonances except for that due to the carbinol carbon a t 77.28 ppm lie between those found for the free racemic acid and its sodium salt. These extra peaks are likely due to ligands that are free or only loosely associated with arsenic(II1) owing to total or partial dissociation of the complexes. That the peaks seem to lie between those of the salt and those of the free acid is not unexpected since the solutions are relatively acidic (apparently owing to formation of moderately soluble As(OH), from relatively acidic and unstablez3 cationic arsenic(II1) species formed initially upon complex dissociation). The intensities of the extra peaks are much higher (relative to those attributed to the binuclear anionic complex) than those found for the arsenic(II1) tartrates and substituted tartrates.' This indicates a much greater dissociation for the arsenic (*)-dhcdc complex, not an unexpected observation since the ligand is known to be in an unfavorable conformation. Apparently the antimony(II1) (f)-dhcdc complex has such a sufficiently increased stability (relative to that of dissociated species) that no dissociation is detected. Since the "C spectrum of the arsenic(II1) (*)-dhcdc complex is complicated by additional resonances, a crystal structure of the sodium salt has been carried out both to confirm the binuclear structure and isom-
~~
(23) Cotton, F. A,: Wilkiinson, 0.'Advaneed Inorganic Chemistry', 3rd ed.; Interscience: New York, 1972; p 392.
3340 Inorganic Chemistry, Vol. 21, No. 9, 1982
Marcovich et al.
Table lV. Interatomic Distances and Angles for Na,[As2(C,H,0,),]~6H,0 with Estimated Standard Deviations in Parentheses Distances, A 2.002 (3) 2.021 (3) 2.044 (3) 1.981 (3) 1.792 (2) 1.800 (2) 1.803 (2) 1.803 (2) 1.306 (4) 1.289 (5) 1.284 (4) 1.299 (5) 1.214 ( 5 ) 1.226 ( 5 ) 1.229 (5) 1.218 ( 5 )
O( l)-As(l)-0(7)
0(5)-As(2)-0(11) 0(3)-As( 1)-0(9) 0(4)-As(2)-0(10) O( l)-A~(l)-0(3) 0(7)-A~(1)-0(9) 0(4)-As(2)-0(5) O( 10)-A~(2)-0(11) C(l)-O( l)-As( 1) C(5)-0(7)-As( 1) C(4)-0(5)-As(2) C(8)-0( ll)-As(2) C(2)-0(3)-A~(l) C(6)-0(9)-As( 1) C(3)-0(4)-A~(2) C(7)-0(10)-A~(2) C(l)-C(2)-0(3) C(4)-C(3)-0(4) C(5)-C(6)-0(9) C(8)-C(7)-0( 10)
C(3)-0(4) C(6)-0(9) C(7)-0(10) C(1)4(2) C(3)-C(4) W)-C(6) C(7)4(8) C(2)-C(3) C(6)-C(7) C(2)-C(9) C(3)-C(12) C(6)-C(13) C(7)-C(16) C(9)-C(lO) C(ll)-C(12) C(13)4(14) C(15)-C(16) C(lO)-C(ll) C(14)-C(15)
1.427 (4) 1.433 (4) 1.431 (4) 1.434 (4) 1.534 (5) 1.534 (5) 1.538 ( 5 ) 1.529 (5) 1.555 (5) 1.548 ( 5 ) 1.536 (5) 1.534 ( 5 ) 1.534 (5) 1.542 (5) 1.522 (6) 1.526 (5) 1.531 (6) 1.533 (6) 1.520 (6) 1.513 (6)
Anglesi, Deg 158.2 (1) C(2)-C(3)C(12) 157.3 (1) C(3)-C(2)-C(9) 104.2 (1) C(6)-C(7)4(16) 104.5 (1) C(7)-C(6)-C(13) 83.7 (1) C(2)4(9)-C(lO) 83.9 (1) C(3)-C(12)-C(ll) 82.8 (1) C(6)4(13)-C(14) 83.9 (1) C(7)-C(16)-C(15) 114.1 (2) C(9)-C(lO)-C(ll) 114.2 (2) C(lO)-C(ll)-C(l2) 113.6 (2) C(13)4(14)-C(15) 114.5 (2) C(14)-C(15)-C(16) 118.8 (2) 0(1)-C(l)-C(2) 119.3 (2) 0(5)-C(4)-C(3) 119.0 (2) 0(7)-C(S)-C(6) 118.0 (2) O(ll)-C(8)-C(7) 109.1 (3) 0 ( 3 ) 4 ( 2 ) 4 ( 3 ) 108.9 (3) 0(4)-C(3)
